AI agents for medication management, working as one workforce.
OneDose is six AI agents that each own one part of the medication journey — inbound calls, prescription intake, refills, prior authorization, adherence and post-discharge follow-up. They share one patient context and one audit log, so the work hands off between them without the patient repeating themselves or the care team reading six histories.
Every agent runs with the same boundary: it does the operational work and routes the clinical decision to a human.
What are the six agents?
Each agent has its own page describing what it does, how it works, and — the part worth reading — what it deliberately will not do.
Follow-up Agent
The schedule is derived from the discharge summary, the prescription and the risk profile, so a complex cardiac discharge and a routine day-case do not get the same calendar.
automated post-discharge follow-upAdherence Agent
Adherence is a behaviour, not an event, so the agent is measured on the pattern across weeks rather than on whether a single reminder was delivered.
AI medication adherence agentVoice Agent
It is a phone number, not an app. The patient does not download anything, log in, or learn a menu tree — they call the pharmacy the way they always have.
AI voice agent for pharmacyIntake Agent
Intake is where the rest of the medication journey gets its data. A wrong strength captured here is a wrong strength in every downstream reminder, refill and check-in.
AI patient intake structuringRefill Agent
The refill queue is the highest-volume, lowest-judgement work in a pharmacy — which is exactly the profile of work worth automating first.
automated prescription refill processingClaims & PA Agent
Prior authorization is mostly a documentation and follow-up problem wearing a clinical costume: the evidence usually exists, and the cost is the hours spent assembling it and chasing the answer.
prior authorization automation
Why does one shared context matter?
Because six tools that do not talk to each other produce a patient who repeats themselves and a care team that reads six histories. That is not a smaller version of a workforce — it is a different, worse thing.
A patient calls about a side-effect. The voice agent already knows the follow-up agent asked about that medication yesterday and that the adherence agent has seen two missed doses since. The escalation that reaches the pharmacist carries all three facts, so the pharmacist opens the call knowing what happened rather than assembling it.
The same property is what makes the audit log worth having. One log across the journey answers “what did we do for this patient, and when?” — six logs answer it six times and agree with each other approximately.
Where is the line between the agents and your clinicians?
In the same place for every agent, which is the point. Each one handles the operational work — the call, the form, the queue, the chase — and routes the clinical decision to a human with the full context attached.
Clinical questions, low-confidence cases, any sign of deterioration and any request for a person all escalate. The agent collects and escalates; the clinician decides. That boundary is a design constraint rather than a configuration setting, because it is the constraint that makes the rest of it deployable in a hospital at all.
Frequently asked
- What are AI agents for medication management?
- AI agents for medication management are specialized software agents that each own one part of the medication journey — answering inbound calls, structuring prescriptions, processing refills, handling prior authorization, tracking adherence, running post-discharge follow-up — and coordinate through one shared patient context, escalating clinical decisions to a pharmacist or clinician.
- How many agents does OneDose have?
- Six: the Voice Agent, the Intake Agent, the Refill Agent, the Claims & PA Agent, the Adherence Agent and the Follow-up Agent. They share one patient context and one audit log, which is what makes them a workforce rather than six separate tools.
- Why is shared context between agents important?
- Because without it the patient repeats themselves and the care team reads six histories. When the voice agent already knows what the follow-up agent asked yesterday, a patient calling about a side-effect they reported the day before does not have to explain it again — and the escalation reaching a pharmacist carries both conversations.
- Do the agents make clinical decisions?
- No. Every agent runs with the same boundary: it handles the operational work and routes clinical questions, low-confidence cases and any sign of deterioration to a human with the full context attached. The agent collects and escalates; the clinician decides.
- Can we deploy only one agent?
- Yes. Most deployments start with one — usually the Voice Agent for a pharmacy with a call-volume problem, or the Follow-up Agent for a hospital focused on discharges. The shared context is what makes adding the second one worth more than the first.
- Does OneDose replace our pharmacy or EMR system?
- No. OneDose is built to read and write against the systems you already run rather than replace them. Integration depth varies by system and region — ask us what is live for your estate and we will confirm it in writing before you commit to anything.